ON THE STRUCTURE, GENESIS AND SIGNIFICANCE OF DNA DUPLICATIONS AT THE ROUS-SARCOMA PROVIRAL INSERTION SITES IN RAT-1 CELLS

We have previously shown that most Rous sarcoma proviruses in Rat-1 cells are simple insertions, at apparently random sites, and are not transcribed. A small minority of simple insertions are transcribed and these show a bias to insertion at sites closely 3' to presumptive cellular CG-rich islands. However, most transcribed proviruses are complex, typified by contiguous duplications of proviral DNA 5' to a complete proviral unit. The cellular sites of these complex insertions are unknown and their structure and significance incompletely documented. We report here more extensive analyses of proviral duplications, with the following findings. The 5' duplications predominantly involve proviral regions known to contain enhancer functions, substantiating earlier data. The cellular insertion sites are not biased to CG-rich loci at the level displayed by simple transcribed proviruses. The detailed structure of two duplications, and partial analysis of several others, strongly favours their genesis by illegitimate template transfer at reverse transcription, followed by self-recombination. These findings show that aberrant reverse transcription can generate duplications that dispense with the need for an expressed provirus to be integrated at a favourable cellular site.